Skip to content

fix(providers): mark image-capable OpenRouter models as vision-capable #482

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
mousamax wants to merge 1 commit into
base: main
Choose a base branch
from
+26 −3
Open

fix(providers): mark image-capable OpenRouter models as vision-capable #482

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
16 changes: 13 additions & 3 deletions coworker/providers/matrix.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -211,16 +211,26 @@ class ModelEntry:
"Llama 4 Maverick · via Fireworks", _AGENTIC, 1_000_000
),
# OpenRouter slugs are lowercase `<lab>/<model>` (checked against their catalog
# 2026-07-25); same labs as above, one key for all of them.
# 2026-07-25); same labs as above, one key for all of them. Kimi K2.6 and Llama 4
# Maverick declare image input (architecture.input_modalities, 2026-08-09); PDFs
# unverified over the compat surface, like Kimi K3 above.
"openrouter:z-ai/glm-5.2": ModelEntry("GLM-5.2 · via OpenRouter", _AGENTIC, 128_000),
"openrouter:moonshotai/kimi-k2.6": ModelEntry(
"Kimi K2.6 · via OpenRouter", _AGENTIC, 256_000
"Kimi K2.6 · via OpenRouter",
ModelCapabilities(
tools=True, vision=True, parallel_tool_calls=True, streaming=True
),
256_000,
),
"openrouter:deepseek/deepseek-v4-pro": ModelEntry(
"DeepSeek V4 Pro · via OpenRouter", _AGENTIC, 128_000
),
"openrouter:meta-llama/llama-4-maverick": ModelEntry(
"Llama 4 Maverick · via OpenRouter", _AGENTIC, 1_000_000
"Llama 4 Maverick · via OpenRouter",
ModelCapabilities(
tools=True, vision=True, parallel_tool_calls=True, streaming=True
),
1_000_000,
),
# Stealth/cloaked alpha (catalog-checked 2026-08-24: 1,048,576 ctx, tool calling).
# These are temporary lab previews — expect the slug to vanish when the lab ships
Expand Down
13 changes: 13 additions & 0 deletions tests/test_providers.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -503,6 +503,19 @@ def test_matrix_answers_capabilities_for_reseller_ids():
assert caps.tools and caps.parallel_tool_calls and caps.streaming


def test_openrouter_vision_matches_declared_input_modalities():
"""Vision must track what OpenRouter declares per model, or the engine swaps in its
non-vision placeholder and the image never reaches a model that can read it."""
for mid in (
"openrouter:moonshotai/kimi-k2.6",
"openrouter:meta-llama/llama-4-maverick",
):
caps = capabilities_for(mid)
assert caps.vision and not caps.pdf, mid
for mid in ("openrouter:z-ai/glm-5.2", "openrouter:deepseek/deepseek-v4-pro"):
assert not capabilities_for(mid).vision, mid


def test_matrix_labels_and_custom_model_fallback():
from coworker.providers.matrix import MATRIX, model_labels

Expand Down